How to Utilize AI to Combat the Pandemic?
-
In the fight against the pandemic, governments, organizations, and individuals are doing their utmost to respond. AI, once a symbol of cutting-edge technology, is now being deployed in the battle against the outbreak.
Even those who typically show little interest in artificial intelligence technology have taken notice of AI during this extraordinary period.
On February 4, the Ministry of Industry and Information Technology issued the "Initiative on Fully Leveraging AI to Combat the Novel Coronavirus Pneumonia Outbreak," calling for the rapid use of AI technology to address gaps in epidemic control and to explore its applications in diagnosing and treating COVID-19 as well as in epidemic prevention and control.
This may be the first government initiative in history to combine AI technology with epidemic response. The comprehensive battle against the novel coronavirus pneumonia in China, due to its unique temporal and spatial context, has also become the first instance of AI being applied to a large-scale public health event.
At this moment, various AI technologies and products are fulfilling their roles in laboratories, hospitals, public transportation hubs, and communities, racing against the outbreak.
Let’s take stock of the roles AI has played during the pandemic and the insights these roles offer to the AI industry, the healthcare sector, and society at large.
Objectively speaking, AI has only played a supporting role in the fight against the pandemic. But perhaps when the outbreak is over, we will ask: After this test, can and should we further advance the intelligentization of public health and medical research? If the answer is yes, how can we learn from this experience to shape the future?
The pandemic is inherently brutal, but humanity has often found a way forward through adversity.
Modern medical history generally holds that the 1918 Spanish flu prompted the establishment of modern public health protection systems. From the performance of AI in the fight against the pandemic, we may begin to reflect on more possibilities.
After the outbreak, one of the first announcements made by major cloud computing companies was the free provision of AI computing power to research and medical institutions.
Many netizens wondered: What does AI computing power have to do with fighting the pandemic?
This brings us to a fundamental aspect of AI computing: processing unstructured data matching through tensor calculations.
In classical computing environments, many unstructured data tasks—such as image recognition, speech synthesis, gene matching, and geological information calculations—cannot be efficiently processed.
This necessitates specialized computing chips and architectures for AI, making AI computing power a key focus for chip and cloud service providers in recent years.
The accumulated industrial capacity in AI computing has proven invaluable during the pandemic, accelerating medical analysis capabilities.
In today’s medical analysis field, tasks such as viral gene sequencing, protein target screening, and matching historical data on viruses and drug development all rely on AI computing power. Better algorithms can significantly improve the efficiency of related tests.
These tasks are foundational for understanding the nature of the virus, developing better treatment plans, and creating vaccines and targeted drugs—their value is self-evident.
While AI computing in virus analysis and vaccine development only shortens matching cycles and improves detection efficiency—rather than autonomously developing vaccines as one might imagine—the availability of ample AI computing power is crucial in the race against time during the pandemic.
Currently, major cloud computing providers have opened their AI computing resources for free to support the fight against the outbreak.
Some tech companies have also made gene detection algorithms freely available to gene testing institutions, epidemic prevention centers, and academia, reducing the time required for COVID-19 gene detection.
In this pandemic response, the rapid isolation of the virus, completion of gene sequencing, and the swift emergence of research information on treatments owe much to the accumulated industrial capabilities in AI computing and algorithms.
From research institutions and laboratories to the front lines of the pandemic, AI has played many critical roles. Among the most valuable for the current phase are robots with visual recognition and voice interaction capabilities, which can replace medical staff in patient care.
The use of medical robots for COVID-19 treatment in the U.S. recently sparked discussions in China.
However, medical robots require relatively mature industrial support and are difficult to deploy quickly.
In China, some have repurposed hotel AI robots to deliver medicines and medical supplies. On the front lines, AI’s most helpful capability is diagnostic assistance, currently focused on medical imaging and AI analysis.
Within a week of the outbreak, many domestic AI companies’ medical imaging systems began operating in major hospitals, providing intelligent systems based on medical imaging analysis for healthcare workers and patients.
While COVID-19 diagnosis primarily relies on reagent testing, patients’ lung images also exhibit identifiable characteristics.
AI technology can compress traditional tests that take hours into seconds. This capability effectively supplements reagent testing, aiding rapid diagnosis and addressing the shortage of medical personnel.
It is foreseeable that AI-assisted diagnostic capabilities based on medical imaging will soon be deployed on the front lines. Major AI companies have urgently enhanced their products in this area and are collaborating more closely with medical research institutions.
With the return of travelers, airports, train stations, and highway checkpoints have become critical points for epidemic control. The long queues for temperature checks have become a necessary inconvenience during this period, but the prolonged waits also create large gatherings, posing additional risks.
In recent days, you may have noticed the introduction of contactless temperature checks in many places, allowing people to pass through without stopping or removing masks.
In such systems, AI is indispensable.
First, AI must accurately identify faces without requiring masks to be removed, matching individuals with their test data.
Second, it requires body recognition and tracking to compare with temperature thresholds from sensors, flagging individuals with abnormal temperatures.
Paired with infrared and visible light sensors, AI temperature screening significantly improves throughput in public spaces.
Based on products deployed in Beijing and elsewhere, a single device can screen 10–20 people per second, matching the normal flow rates of train stations, airports, and subway entrances.
In public spaces, AI also contributes to epidemic control through public security applications.
For example, AI cameras can track individuals’ movements in public areas using facial recognition. This technology has not only improved public safety in recent years but has also proven highly effective during the pandemic.
For instance, it has traced transmission chains involving mere seconds of contact, which the individuals involved were unaware of, preventing unpredictable large-scale spread.
The integration of public health security and AI technology is redefining the balance between safety and efficiency. The AI applications during this pandemic are likely to influence the long-term development of public health security systems.
Another invisible aspect of epidemic control occurs over the phone.
If you’ve traveled during the outbreak, you may have received an AI call asking about your trip, including dates, locations, and your health upon return. These data, collected through AI calling systems, form the foundation for grassroots epidemic control.
Given China’s vast population and the scale of travel during the Spring Festival, community-level census and notification efforts have become immense challenges. With limited staff and overwhelming tasks, around-the-clock manual calls are impractical.
Thus, the repetitive task of calling has become an unbearable yet essential burden for grassroots workers. In this context, mature intelligent customer service systems have emerged as a solution.
AI customer service systems, based on smart calling and voice interaction, can quickly transform into intelligent investigators, handling tasks like population screening, follow-ups, and notifications with hundreds of times the efficiency of manual calls.
Going further, some intelligent phone systems can perform relatively complex epidemic prevention tasks, such as random surveys of users to collect data on their living conditions and health status, forming statistical samples. They can also conduct continuous follow-ups with key populations to establish a targeted prevention and control system.
At the same time, medical management units and grassroots organizations can leverage the currently free and open intelligent phone capabilities to develop more tailored systems, allowing AI to become part of social care and connectivity during special times.
Reviewing the integration of AI and epidemic prevention efforts, it is clear that, unlike other technologies, AI can penetrate the core of various levels of epidemic response. For instance, while the internet primarily serves as a communication tool—an irreplaceable and critical role—it cannot directly accelerate virus analysis or vaccine development like AI can.
AI can cover core tasks across various fields, a realization that has emerged during this epidemic as a prerequisite for society's understanding of AI technology. However, it is important to note that AI's fundamental capability lies in improving industrial efficiency and replacing repetitive tasks, not substituting human labor entirely.
In other words, AI in epidemic prevention is an auxiliary tool—albeit a crucial one—and an accelerator. Overall, AI can play a role in epidemic control under three conditions:
These three characteristics of AI's role have been reiterated many times. The reason for emphasizing them again here is to call on AI developers, as well as those in medical products, genetics, and robotics, to refocus on AI's foundational capabilities and collaborate to identify unique scenarios where their expertise can be applied.
Currently, major AI companies and cloud computing providers are rapidly deploying AI solutions for epidemic prevention, offering them for free. However, the coverage from a few leading companies is insufficient. Broader developer participation is needed to maximize AI's value and advance to the 2.0 stage of AI-driven epidemic response.
Platform-based AI companies are now not only contributing their own capabilities but also empowering developers with open tools and models to address long-tail needs and improve overall efficiency in the fight against the epidemic.
This process requires collaboration and efficient communication among internet companies, AI developers, and medical and scientific researchers. Many AI developers are eager to contribute but lack familiarity with medical scenarios, data, and standards. Greater involvement from medical professionals in designing these solutions would also be a significant contribution.
This is not the time to praise AI's role in epidemic prevention, as the battle is far from over. While AI is not the protagonist, its auxiliary functions in key areas offer glimpses of future possibilities. Ideally, AI could one day take a leading role in public health protection, reducing the burden on healthcare workers.
This marks AI's first large-scale deployment in epidemic response. Although AI companies have demonstrated rapid responsiveness and social responsibility, the broader societal and medical systems' adoption and integration of AI remain at an early stage. For example:
The swift deployment of AI in this crisis relied heavily on the maturity of cloud and AI infrastructure, which is now largely ready for immediate use. However, the AI capabilities available to frontline workers are still limited, with insufficient scenario coverage and many solutions only applicable in major cities, unable to reach critical areas like Wuhan promptly. This highlights the need for deeper industry integration.
AI can be seen as a combination of foundational capabilities and algorithms that can be quickly assembled into specific products and platforms. Yet, during this epidemic, medical professionals, researchers, and tech developers often lacked familiarity with these tools, requiring extensive communication and delaying nationwide deployment.
While AI has performed well in software applications, its hardware presence—such as medical robots equipped with remote controls, cameras, microphones, or stethoscopes—has been notably absent from the frontlines. Bridging the gap between software AI and AI robotics requires a robust AI + IoT industrial chain. The current weakness in hardware ecosystems hampers the expected efficiency gains across industries, a challenge that must be addressed in the next technological cycle.
AI has contributed significantly to epidemic response, but could it do more? The answer is yes.
For instance, some U.S. AI startups have used social media data to predict outbreaks. Could this become a future tool to provide early warnings and prevent critical signals from being lost in the noise of social media?
Similarly, intelligent management of public health systems and smart allocation of emergency supplies are areas with mature solutions. However, these require long-term preparation and were not visible in this outbreak. For example, recent controversies over delayed resource allocation due to staffing shortages raise the question: Could AI automate this process? Such solutions are common in smart industry and logistics but have yet to be deployed in epidemic response.
The severity of the epidemic has left a deep impression on everyone. There is no doubt that we will overcome it, and soon. But we must also consider what legacy it leaves behind.
When the crisis passes, we should ask ourselves: What lessons can we take forward? These questions extend beyond AI, healthcare, or this particular outbreak—but reflection and action are always timely.